feat(cuda): add CUDA IPC tensor transport + Python data-scientist ergonomics#127

Open

YuanYuYuan wants to merge 20 commits into

mainfrom

Commits on Mar 13, 2026

feat(cuda): CUDA tensor transport via ZSliceKind::CudaPtr
YuanYuYuan
committed
feat(nix): add CUDA dev shells for zenoh-cuda feature
YuanYuYuan
committed
chore(cuda): patch full zenoh stack for cuda-zslice feature branch
YuanYuYuan
committed
feat(nix): expose host NVIDIA driver in cuda dev shell
YuanYuYuan
committed
fix(nix): set CUDA_PATH to cuda_cudart.lib (has libcudart.so)
YuanYuYuan
committed
feat(cuda): publisher API — ZBuf::from_cuda, PyCudaBuf, cuda_pubsub example
YuanYuYuan
committed
feat(cuda): cuda_slices() helper, from_device_ptr binding, integration tests
YuanYuYuan
committed
fix(cuda): enable SHM transport in with_shm_enabled(); fix cuda_pubsub.py
YuanYuYuan
committed
fix(cuda): fix DLPack double-free; add --warmup to cuda_pubsub.py
YuanYuYuan
committed
feat(cuda): typed tensor transport — TensorMeta, from_torch(), CudaTensor ZSlice
YuanYuYuan
committed
feat(ros-z-py): data-scientist ergonomics — as_torch, as_numpy, publish_tensor, pyi stubs
YuanYuYuan
committed
fix(ros-z-py): clippy — wrong_self_convention allow, remove redundant Ok wrap
YuanYuYuan
committed
docs(cuda): book chapter gpu-tensor-transport; update example to publish_tensor/as_torch
YuanYuYuan
committed
chore(deps): update Cargo.lock for zenoh feat/mem-transport
YuanYuYuan
committed
feat(cuda): ORT cross-language ML inference pipeline demo
YuanYuYuan
committed
fix(deps): replace local zenoh path patches with git refs to ZS feat/mem-transport
YuanYuYuan
committed
fix(rmw): serialize_to_shm<B> — match generic ShmProviderBackend trait signature
YuanYuYuan
committed
test(cuda): typed tensor metadata and native IPC handle tests
YuanYuYuan
committed
chore(ci): drop FlakeHub cache — use Cachix only
YuanYuYuan
committed
trigger CI
YuanYuYuan
committed